RPi Datadog Temperature Collector

A docker container for collecting metrics from a GPIO temperature sensor to be sent to Datadog. This was intended to be used for homebrewing, but can be used for anything where real time temperature data is desired.

Prerequisites

Hardware setup

First, we need to connect the thermometer to the Raspberry pi and load the drivers. I used modmypi's tutorial to get this set up.

Create a free tier Datadog account

You should be able to keep these metrics with 1 day of retention with a free-tier datadog account. Create an account, and add an app key and an api key for our client to authenticate.

Install Docker on your Raspberry Pi

Install the latest version of docker

Usage

Pull the code to your raspberry pi.

git clone git@github.com:cgpuglie/rpi-datadog-temperature

Create a docker-compose file in the repo directory with your desired configuration. For example:

version: '3.1'
services:
  pi-dd-temp:
    image: pi-dd-temp
    privileged: true
    restart: always
    build: .
    devices:
      - /dev/gpiomem:/dev/gpiomem
    volumes:
      - /lib/modules:/lib/modules
    environment:
      DD_METRIC_NAME: 'belgian-strong.temperature'
      DD_API_KEY: 'my-api-key'
      DD_APP_KEY: 'my-app-key'
      DD_ENDPOINT: 'api.datadoghq.com'
      GRANULARITY: 20000
      DEVICE: '28-0417a2ef0dff'

These options are available for configuring the behavior of the client. The values shown are default.

# Datadog configuration
## Metric name, how it will appear in datadog
DD_METRIC_NAME='temperature',
# Authentication info
DD_API_KEY='',
DD_APP_KEY='',
DD_ENDPOINT='api.datadoghq.com',
# App configuration
## How frequently to pull for temperature, in ms
GRANULARITY=20000,
## Name of the device, found with `cd /sys/bus/w1/devices/; ls`
## See tutorial for setting up thermometer for more information
DEVICE=''

Build the image.

docker-compose build

Note: This may take a little bit due to i/o performance on the raspberry pi. I have considered mounting /var/lib/docker to a usb drive or tmpfs, but haven't experimented with this. If you understand the implications and want to try it, feel free.

Run the image in the background.

docker-compose up -d

That's it. You should see metrics showing up in Datadog.

cgpuglie / rpi-datadog-temperature